the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Mar 2023
Evaluation of TROPOMI operational standard NO2 column retrievals (from version 1.3 to 2.4) with OMNO2 and QA4ECV OMI observations over China
Abstract. The TROPOMI satellite instrument plays a key role in nitrogen dioxide (NO2) monitoring on account of its unprecedented spatial resolution and stable quality of data. However, since 2019, TROPOMI operational NO2 retrieval has improved and updated in three versions (1.4, 2.2 and 2.4), with significant impact on retrieved NO2 column. Thus, studies including both TROPOMI NO2 data before and after the activation of these versions could show artificial jumps. Moreover, up to date evaluation result of TROPOMI NO2 data in current version 2.4 is not yet well documented in the literature. Therefore, in this work, we focus on evaluating TROPOMI's capability to detect NO2 under the different retrieval version conditions, by comparing with OMNO2 data and QA4ECV OMI data over China. We find a 38 % increase of tropospheric NO2 in version 1.4 due to improved FRESCO-wide cloud retrieval, and a 14 % increase in version 2.2 due to adjusted surface albedo for cloud-free scenes. We show that the upgrade to version 2.4 with new DLER surface albedo, led to an increase by 3 x 1014 molecules cm-2 of tropospheric NO2 over vegetation. Furthermore, we demonstrate that TROPOMI data shows strongest tropospheric NO2 seasonal variation compared to OMNO2 data and QA4ECV OMI data, and this seasonal effect was enhanced with the tropospheric NO2 retrieval version upgrades. Additionally, we examine for the first time the change of TROPOMI AMFs (air mass factors) in the different versions, and based on it, we arrive at a correction for the underestimation of TROPOMI NO2 column in previous versions. We also find a 33 % overestimation of NO2 reduction during the COVID-19 lockdown over China when using TROPOMI data before and after the activation of the NO2 version 1.4.
- Preprint
(2136 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2023-175', Anonymous Referee #1, 02 May 2023
In their paper Jianbin Gu and co-authors evaluate the currently available versions of the operational TROPOMI NO2 retrieval and compare these with two OMI NO2 products over China. After comparing the results with the existing literature (a good list of references is contained in the paper) and documentation for TROPOMI NO2 I concluded that there is not much new information contained in the paper. As such I am not in favour of publishing these results.
Let me illustrate my conclusion by commenting the abstract:
- l 27: "We find a 38 % increase of tropospheric NO2 in version 1.4 due to improved FRESCO-wide cloud retrieval, and a 14 % increase in version 2.2 due to adjusted surface albedo for cloud-free scenes."
This result is very similar to the conclusion drawn in paper of van Geffen, 2022, where the retrieval changes in version 2.2 (and 1.4) are discussed. Also the comparison against OMI-QA4ECV are presented in this paper.- l 29: "We show that the upgrade to version 2.4 with new DLER surface albedo, led to an increase by 3 x 1014 molecules cm-2 of tropospheric NO2 over vegetation."
The paper includes only two months of data of version 2.4, which I would judge is not enough to document the impact of the upgrade to v2.4. Possible weather influences are not discussed by the authors. The increase of NO2 over vegetation has been discussed in the release documentation, the ATBD and the readme file. As indicated below, I was not convinced by the analysis on this topic.- l 31: "we demonstrate that TROPOMI data shows strongest tropospheric NO2 seasonal variation compared to OMNO2 data and QA4ECV OMI data, and this seasonal effect was enhanced with the tropospheric NO2 retrieval version upgrades": In the paper by van Geffen it was also reported that the largest increases occur in wintertime, so not really a new results.
- l 33: "we arrive at a correction for the underestimation of TROPOMI NO2 column in previous versions"
In this respect the paper is a bit late. Such corrections are no longer relevant given the S5P-PAL (available since december 2021) and the reprocessing of v2.4 (available since March 2023). Such corrections have been introduced before, e.g. Riess et al, 2021.- l 35: "We also find a 33 % overestimation of NO2 reduction during the COVID-19 lockdown over China when using TROPOMI data before and after the activation of the NO2 version 1.4." This follows directly from the large increase introduced in v1.4 and is a bit a trivial result. A better analysis of the COVID period was a main motivation to launch the S5P-PAL reprocessing.
Further comments:
- Why do the authors focus only on China? The author team is from China, but the paper is submitted to an international journal. It would be just as relevant to know the impacts over Europe, USA, the tropics etc. As indicated by Van Geffen et al., 2022, the impact seems to be quite dependent on the region.
- Why is a reference to the S5P-PAL dataset (https://data-portal.s5p-pal.com/products/no2.html) not included? This PAL dataset was generated to remove the jumps between versions, e.g. for Covid studies.
- There is a new reprocessing available since March 2023, covering the full mission duration, see
https://sentinels.copernicus.eu/web/sentinel/-/copernicus-sentinel-5-precursor-full-mission-reprocessed-datasets-further-products-release .
The seasonality linked to the DLER albedo update can be studied with this new dataset. I realise that the paper was written before the reanalysis became available, so this is not a main reason for my negative judgement.- The paper focusses on China, but the POMINO product is not mentioned at all
(http://www.pku-atmos-acm.org/acmProduct.php/#TROPOMI). This is a clear omission.- It was surprising that Lamsal et al, https://doi.org/10.5194/amt-14-455-2021, is not cited for v4 of OMNO2A.
- There is no reference to the product readme file and user manual documents of TROPOMI NO2: these documents inform users of the updates of the processor and main impacts and are therefore relevant for this paper.
- As indicated by the authors, the v2.4 upgrade is "not well documented" in the peer-reviewed literature. I can sympathise with this statement. But this version is new, and is also used in the recent reprocessing. Did the authors get in contact with the retrieval team, which would be a normal step to take?
- The analyses presented by the authors are rather straightforward, consisting of simple comparisons of averages of the tropospheric column and the tropospheric air-mass factor between different retrieval versions. A more in-depth analysis of the differences is missing, and the general conclusions broadly agree with what has already been reported before.
- The data series stops in September 2022. This is a very short period (only two months) to make any statements on the version 2.4 data. Because the albedo climatology is available on a monthly basis, it would be important to document a full year of data.
- I found the analysis of the impact over vegetation not convincing: How can we compare relative differences in Fujian and the entire China? Many aspects may play a role here. Furthermore, the analysis is limited to one month, which is not convincing as weather variability may play a big role. So I do not think the authors have presented enough evidence to quantify the increase due to v2.4.
- Concerning the seasonality, section 3.3: I could not reconcile the results of Figure 5 with figure 6, for QA4ECV OMI. Is there a mistake in figure 5, since the seasonality seems much too small?
Citation: https://doi.org/10.5194/egusphere-2023-175-RC1 -
AC1: 'Reply on RC1', Jianbin Gu, 30 May 2023
We sincerely appreciate the Reviewer for examining our manuscript and correctly pointing out that some areas of the manuscript required further improvement. We have conducted a thoroughly revision according to the Reviewer’ comments and suggestions. Please find a detailed point-by-point response to the Reviewer’ comments in the supplement.
In the response letter, comments by the Reviewer are italicized and in blue for clarity.
Many thanks for examining our manuscript.
-
AC1: 'Reply on RC1', Jianbin Gu, 30 May 2023
-
RC2: 'Comment on egusphere-2023-175', Anonymous Referee #2, 04 May 2023
The paper presented by Gu et al. evaluated the TROPOMI operational standard NO2 column under the different retrievals by comparing with OMNO2 and QA4ECV OMI observations over China. It falls into the scope of AMT journal and well written. However, there some concerns need to be addressed before it can be considered for publication.
- Line 123-124, the definition of SCD and VCD need be rephrased since they are not accurate in current description. Please refer to the product documentation or the DOAS book (Platt and Stutz, 2008).
- In Fig.2, in addition to the version of retrievals, the periods are also different. How can we attribute the differences to the retrieval itself rather than the differences of NO2 in temporal? Please clarify and provide the evidence.
- Sect 3.2, when the authors discuss the impacts of DLER over vegetation, only a summer month (August) were selected for Fujian province and China. I guess this month was chose to represent the condition with vivial vegetation. However, a comparative withered season/month should also be considered to show the change in surface albedo and further impacts in DLER and NO2 products.
- Fig. 5 and other similar inferred conclusions, I think that an independent ground-based measurements of NO2 VCD datasets can strongly enhance the evidences. Otherwise, it’s hard to exclude the upward trends in winter and downward in summer from the seasonal pattern difference from year to year. Similarly, there also other conclusions are not solid and convincible.
- Line 426-428, how to create the AMF dataset? By RTM? If it is, please describe the simulation and key inputs in details. And how about the authors’ simulation compared to the AMFs used in products retrieval? If not, how to get the AMF?
- Line 487-503, considering there were many literatures that reported the changes of NO2 VCDs during the 2020 lockdown in China (in both spaced-based sensors and ground based MAX-DOAS), the authors could refer to the reported decreases and compared with the expectation in Figure. 9.
Minors:
- Better to cite the full name of some nouns for the first time even in the abstract, e.g. Tropomi, OMI, QA4ECV, DLER, etc.
- Line 165, NO2 subscript.
- Line 193, should be “tropospheric and stratospheric NO2 SCDs”
- I would like to suggest to show the monthly series of different products of NO2 VCDs from OMI and TROPOMI in another panel in Figure 1 too, which is helpful to show the absolute differences.
- Line 328-330, the comparison of spatial distribution between annual averages of these three products may be also presented in Fig. 3.
Citation: https://doi.org/10.5194/egusphere-2023-175-RC2 -
AC2: 'Reply on RC2', Jianbin Gu, 30 May 2023
We sincerely appreciate the Reviewer for examining our manuscript and correctly pointing out that some areas of the manuscript required further improvement. We have conducted a thoroughly revision according to the Reviewer’ comments and suggestions. Please find a detailed point-by-point response to the Reviewer’ comments in the supplement.
In the response letter, comments by the Reviewer are italicized and in blue for clarity.
Many thanks for examining our manuscript.
Status: closed
-
RC1: 'Comment on egusphere-2023-175', Anonymous Referee #1, 02 May 2023
In their paper Jianbin Gu and co-authors evaluate the currently available versions of the operational TROPOMI NO2 retrieval and compare these with two OMI NO2 products over China. After comparing the results with the existing literature (a good list of references is contained in the paper) and documentation for TROPOMI NO2 I concluded that there is not much new information contained in the paper. As such I am not in favour of publishing these results.
Let me illustrate my conclusion by commenting the abstract:
- l 27: "We find a 38 % increase of tropospheric NO2 in version 1.4 due to improved FRESCO-wide cloud retrieval, and a 14 % increase in version 2.2 due to adjusted surface albedo for cloud-free scenes."
This result is very similar to the conclusion drawn in paper of van Geffen, 2022, where the retrieval changes in version 2.2 (and 1.4) are discussed. Also the comparison against OMI-QA4ECV are presented in this paper.- l 29: "We show that the upgrade to version 2.4 with new DLER surface albedo, led to an increase by 3 x 1014 molecules cm-2 of tropospheric NO2 over vegetation."
The paper includes only two months of data of version 2.4, which I would judge is not enough to document the impact of the upgrade to v2.4. Possible weather influences are not discussed by the authors. The increase of NO2 over vegetation has been discussed in the release documentation, the ATBD and the readme file. As indicated below, I was not convinced by the analysis on this topic.- l 31: "we demonstrate that TROPOMI data shows strongest tropospheric NO2 seasonal variation compared to OMNO2 data and QA4ECV OMI data, and this seasonal effect was enhanced with the tropospheric NO2 retrieval version upgrades": In the paper by van Geffen it was also reported that the largest increases occur in wintertime, so not really a new results.
- l 33: "we arrive at a correction for the underestimation of TROPOMI NO2 column in previous versions"
In this respect the paper is a bit late. Such corrections are no longer relevant given the S5P-PAL (available since december 2021) and the reprocessing of v2.4 (available since March 2023). Such corrections have been introduced before, e.g. Riess et al, 2021.- l 35: "We also find a 33 % overestimation of NO2 reduction during the COVID-19 lockdown over China when using TROPOMI data before and after the activation of the NO2 version 1.4." This follows directly from the large increase introduced in v1.4 and is a bit a trivial result. A better analysis of the COVID period was a main motivation to launch the S5P-PAL reprocessing.
Further comments:
- Why do the authors focus only on China? The author team is from China, but the paper is submitted to an international journal. It would be just as relevant to know the impacts over Europe, USA, the tropics etc. As indicated by Van Geffen et al., 2022, the impact seems to be quite dependent on the region.
- Why is a reference to the S5P-PAL dataset (https://data-portal.s5p-pal.com/products/no2.html) not included? This PAL dataset was generated to remove the jumps between versions, e.g. for Covid studies.
- There is a new reprocessing available since March 2023, covering the full mission duration, see
https://sentinels.copernicus.eu/web/sentinel/-/copernicus-sentinel-5-precursor-full-mission-reprocessed-datasets-further-products-release .
The seasonality linked to the DLER albedo update can be studied with this new dataset. I realise that the paper was written before the reanalysis became available, so this is not a main reason for my negative judgement.- The paper focusses on China, but the POMINO product is not mentioned at all
(http://www.pku-atmos-acm.org/acmProduct.php/#TROPOMI). This is a clear omission.- It was surprising that Lamsal et al, https://doi.org/10.5194/amt-14-455-2021, is not cited for v4 of OMNO2A.
- There is no reference to the product readme file and user manual documents of TROPOMI NO2: these documents inform users of the updates of the processor and main impacts and are therefore relevant for this paper.
- As indicated by the authors, the v2.4 upgrade is "not well documented" in the peer-reviewed literature. I can sympathise with this statement. But this version is new, and is also used in the recent reprocessing. Did the authors get in contact with the retrieval team, which would be a normal step to take?
- The analyses presented by the authors are rather straightforward, consisting of simple comparisons of averages of the tropospheric column and the tropospheric air-mass factor between different retrieval versions. A more in-depth analysis of the differences is missing, and the general conclusions broadly agree with what has already been reported before.
- The data series stops in September 2022. This is a very short period (only two months) to make any statements on the version 2.4 data. Because the albedo climatology is available on a monthly basis, it would be important to document a full year of data.
- I found the analysis of the impact over vegetation not convincing: How can we compare relative differences in Fujian and the entire China? Many aspects may play a role here. Furthermore, the analysis is limited to one month, which is not convincing as weather variability may play a big role. So I do not think the authors have presented enough evidence to quantify the increase due to v2.4.
- Concerning the seasonality, section 3.3: I could not reconcile the results of Figure 5 with figure 6, for QA4ECV OMI. Is there a mistake in figure 5, since the seasonality seems much too small?
Citation: https://doi.org/10.5194/egusphere-2023-175-RC1 -
AC1: 'Reply on RC1', Jianbin Gu, 30 May 2023
We sincerely appreciate the Reviewer for examining our manuscript and correctly pointing out that some areas of the manuscript required further improvement. We have conducted a thoroughly revision according to the Reviewer’ comments and suggestions. Please find a detailed point-by-point response to the Reviewer’ comments in the supplement.
In the response letter, comments by the Reviewer are italicized and in blue for clarity.
Many thanks for examining our manuscript.
-
AC1: 'Reply on RC1', Jianbin Gu, 30 May 2023
-
RC2: 'Comment on egusphere-2023-175', Anonymous Referee #2, 04 May 2023
The paper presented by Gu et al. evaluated the TROPOMI operational standard NO2 column under the different retrievals by comparing with OMNO2 and QA4ECV OMI observations over China. It falls into the scope of AMT journal and well written. However, there some concerns need to be addressed before it can be considered for publication.
- Line 123-124, the definition of SCD and VCD need be rephrased since they are not accurate in current description. Please refer to the product documentation or the DOAS book (Platt and Stutz, 2008).
- In Fig.2, in addition to the version of retrievals, the periods are also different. How can we attribute the differences to the retrieval itself rather than the differences of NO2 in temporal? Please clarify and provide the evidence.
- Sect 3.2, when the authors discuss the impacts of DLER over vegetation, only a summer month (August) were selected for Fujian province and China. I guess this month was chose to represent the condition with vivial vegetation. However, a comparative withered season/month should also be considered to show the change in surface albedo and further impacts in DLER and NO2 products.
- Fig. 5 and other similar inferred conclusions, I think that an independent ground-based measurements of NO2 VCD datasets can strongly enhance the evidences. Otherwise, it’s hard to exclude the upward trends in winter and downward in summer from the seasonal pattern difference from year to year. Similarly, there also other conclusions are not solid and convincible.
- Line 426-428, how to create the AMF dataset? By RTM? If it is, please describe the simulation and key inputs in details. And how about the authors’ simulation compared to the AMFs used in products retrieval? If not, how to get the AMF?
- Line 487-503, considering there were many literatures that reported the changes of NO2 VCDs during the 2020 lockdown in China (in both spaced-based sensors and ground based MAX-DOAS), the authors could refer to the reported decreases and compared with the expectation in Figure. 9.
Minors:
- Better to cite the full name of some nouns for the first time even in the abstract, e.g. Tropomi, OMI, QA4ECV, DLER, etc.
- Line 165, NO2 subscript.
- Line 193, should be “tropospheric and stratospheric NO2 SCDs”
- I would like to suggest to show the monthly series of different products of NO2 VCDs from OMI and TROPOMI in another panel in Figure 1 too, which is helpful to show the absolute differences.
- Line 328-330, the comparison of spatial distribution between annual averages of these three products may be also presented in Fig. 3.
Citation: https://doi.org/10.5194/egusphere-2023-175-RC2 -
AC2: 'Reply on RC2', Jianbin Gu, 30 May 2023
We sincerely appreciate the Reviewer for examining our manuscript and correctly pointing out that some areas of the manuscript required further improvement. We have conducted a thoroughly revision according to the Reviewer’ comments and suggestions. Please find a detailed point-by-point response to the Reviewer’ comments in the supplement.
In the response letter, comments by the Reviewer are italicized and in blue for clarity.
Many thanks for examining our manuscript.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 256 | 156 | 11 | 423 | 6 | 9 |
- HTML: 256
- PDF: 156
- XML: 11
- Total: 423
- BibTeX: 6
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1